Summary This proposal addresses two fundamental questions: (1) How do neurons extract features of mechanosensory stimuli that are relevant for motor control? (2) How do central circuits create a flexible linkage between mechanosensory stimuli and behavior? These questions are relevant to human health because sensory processing and sensory-motor integration are disrupted in many neurological and psychiatric disorders. However, sensory processing and sensory-motor integration are not fully understood at the level of cellular mechanisms ? i.e., at the level of neural connectivity, cellular physiology, and synaptic physiology. This level of mechanistic explanation is important to understanding why disease-linked genes produce their characteristic phenotypes. It is also important to developing better therapeutics. As a model system for gaining mechanistic insight into these brain functions, this project will focus on the largest mechanosensory organ in Drosophila (Johnston's organ) and the circuits and behaviors downstream from this organ. Johnston's organ neurons (JONs) encode deflections of the distal antennal segment. These deflections can result from an object touching the antenna, wind, postural changes, or sound. In essence, therefore, Johnston's organ has a range of functions ? somatosensory, vestibular, and auditory. Different JON stimuli elicit different behaviors. These behaviors are variable and context-dependent (not stereotyped action patterns) and so we can use this system to study flexibility in sensory-motor coupling. Our first aim is to determine how JONs encode mechanical stimuli. To test the hypothesis that JONs are highly specialized for specific spatiotemporal features of antennal deflections, we will use a combination of in vivo calcium imaging, electrophysiology, and voltage imaging. Second, we will use in vivo whole cell recordings to test the hypothesis that central neurons postsynaptic to JONs can extract specific frequencies of antennal vibrations by virtue of their intrinsic electrical bandpass filtering characteristics. Third, we will perform in vivo whole cell recordings to investigate how mechanosensory signals are encoded at the level of third-order neurons, and how these signals are relayed to motor control centers. We hypothesize that wind and sound stimuli will be encoded by largely distinct neural channels. Fourth, we will combine whole-cell recording with simultaneous behavioral measurements to determine how mechanosensory cues from JONs steer walking direction in a context-dependent manner. We hypothesize that heading direction cues and context cues will converge at the level of descending motor control neurons that project to the ventral nerve cord. As a whole, this work will provide new insights into the neural computations that occur in mechanosensory processing and mechanosensory- motor integration, as well as the cellular mechanisms that implement these computations.